Research On Electric Vehicle Energy Management Based On Cyber-Physical Computing Technology

With the intensification of global warming,reducing emissions becomes a common issue of the world.In developed countries,the emissions of transportation system are about 1/3 of the total emissions.In China,the transportation system emissions are about 10%of the total.Therefore,promoting electric vehicles(EV)is becoming an important way to reduce emis-sions,due to their low emissions and low fuel consumption.However,because of the charg-ing difficulties,short recharge mileage and inadequate power grid capacity etc.,the promotion of EVs is difficult.To solve these problems,the thesis gives an integrated architecture for the EV energy management system first.Then,based on the architecture,three specific issues are deeply studied,i.e.(1)Charging station deployment,(2)EV charging schedule and control,(3)EV driving planning.To be specific,the main contributions of this paper are shown as follows:First,we study the EV energy management architecture from the perspective of cyber-physical system(CPS).In this part,a common CPS Architecture Description Language(CPSADL)is proposed and the corresponding compiler is designed.By CPSADL,we give the architecture model of EV Energy Management Cyber-Physical Systems(EM-CPS).Second,we study how to size and locate charging stations in traffic networks considering grid constraints to balance the charging demand and power network stability.First,a spa-tio-temporal model of charging demand and a linearized power network model(LPNM)are proposed.Based on the two models,a heuristic algorithm involving the grid constraints(HAG)is designed.The simulation shows that the relative error of the voltage deviation esti-mated by LPNM is about 4%.Compared to the plain demand model,adopting the spa-tio-temporal charging demand model improves the utilization of chargers by 5%at least.Compared to the greedy algorithm with grid constraints(GAG),HAG improves the carrying capacity of the power network by 20.7%,reduces the voltage deviation by 25%and increases the EVs charged by 18.07%.Third,we study EV charging schedule and control by two steps.At the first step,we as-sign the EVs to appropriate charging stations.We model the object as a social warfare model and propose a(2(10)?)approximation algorithm to optimize the assignment.The simulation results show that the proposed algorithm improves the social warfare significantly.At the se-cond step,we focus on the charging power control in one charging station.We propose a common charging management architecture first.Then,based the common architecture,we study the charging power control for the EVs of datacenters'employees.We propose a Val-leyFill scheme and evaluate the scheme by simulations.The results show that ValleyFill not only meets the charging demand,but also reduces the charging cost by 6%.Fourth,we study the driving planning for BEV and PHEV to optimize the energy con-sumption respectively.We propose a dynamic programming algorithm to solve the problem that an optimal route may contain circles.For PHEV,we propose a cost-optimal algorithm(COA)to solve the problem caused by multiple charging modes of PHEVs.The time com-plexity of COA is proved to beO((E~*)~2||~2).The simulations based on real-world map and data show that the average detour caused by COA is 14%.Compared to the shortest path algorithm,COA can reduce the energy cost by 48%.To sum up,this paper combines the CPS with EV energy management system and pro-poses the EM-CPS architecture.Based on the EM-CPS architecture,we study the modeling,analysis and optimization for EV energy management.In detail,the study includes EV charg-ing schedule and control,charging facility deployment and EV driving planning.The results show that the proposed models and methods can benefit the energy saving,emissions reduc-ing etc.significantly.